Variations in paper properties can be due to process conditions but also to variations in wood properties. Better knowledge about the relationships between wood properties and properties of mechanical pulp could lead to a more efficient use of wood resources. AFOCEL has used its thermomechanical pulping (TMP) laboratory pilot refiner to pulp small wood samples and determine the influence wood fibre properties have on pulp fibres. This two-stage laboratory procedure was used within the EU project EuroFiber, which aims at the definition of wood assortments better adapted to the end-product qualities of different European paper mills with TMP plants.

Picea abies (L.) H. Karst. (Norway spruce) trees were sampled in four European countries (Estonia, France, Norway, and Sweden). Samples were taken at different tree heights and split into juvenile, transition, and mature woods, resulting in a total of 450 samples. Wood samples were mechanically pulped after a two-stage refining procedure. Each pulp was characterised by its physical and optical properties, fibre morphology, and specific energy consumption. Data on forest, wood, and pulp properties were compiled and statistically analysed.

Principal Component Analysis revealed a correlation between pulp properties and fibre properties (light-scattering coefficient, fibre length, fines, freeness, brightness, energy consumption).

Five classes of wood were created based on tracheid length and wood density to identify the influence these parameters may have on pulp production (energy consumption, light-scattering coefficient, and tear index).

Finally, a modelling approach enabled us to detect and quantify wood properties and process effects on measured pulp properties. Some categorical data were also tested: wood age (juvenile, transition, and mature wood), country where the tree was grown, and the previous five classes of wood. Part of variances explained by the models was quite low, but showed a general relationship with the specific energy applied during refining and, to a lesser extent, some tendencies with wood properties. With a process that preserved fibre properties well, this work allowed the identification of some wood properties that may be investigated to produce tailor-made pulp.

The hypothesis that lumber value increases when the sawpattern embodies partial knowledge of the internal defect structure was tested. Two methods for estimating the internal structure were evaluated: one used annual ringcounts and the other used the definition of defect core. The former method was applied to pruned Pseudotsuga menziesii (Mirb.) Franco (Douglas fir) logs from the United States and the latter to logs from New Zealand that, although with similar diameters to the United States logs, were substantially younger, pruned much earlier, and had smaller defect cores. Digital models of the logs were constructed. External shape representations were based on direct log measurements. Internal defects were computer-generated for the United States logs, whereas actual branch stub measurements were used for the New Zealand logs. The models were sawn in the AUTOSAW sawing simulator, and the resultant lumber was graded and priced. The sum of lumber prices established a value for each log. That value was compared with lower and upper bounds derived from parallel simulations based on volume-optimising and value-optimising sawpatterns, respectively. The lower bound assumed no a priori knowledge of internal defects while the upper assumed full a priori knowledge.

With the annual ring count method there was no significant difference in mean log value when compared to the lower bound; however, potential to increase value, demonstrated by the difference between lower and upper bounds was significant at 3%. With the defect core method a significant increase of 5% was found and potential to increase value was nearly 11%.

Matching timber quality with end-user requirements is a major research issue and lack of straightness in timber is the most frequent complaint worldwide. The final distortion of timber boards is caused mostly by moisture-related deformations and growth stresses that develop during growth of the tree, but how much the growth rate and growth stresses affect the final shape stability is not fully understood. A finite element analysis in which stress development during tree growth was simulated was performed with the aim of better understanding how growth stresses are generated. The tree growth model was formulated in terms of large strain settings (large changes in volume), whereas the material model for stress development was based on the theory of small strains. An earlier three-dimensional distortion model was developed further for studying the influence of growth stresses on final distortion of the board. The results showed that growth stresses clearly vary during tree growth and they also form a large stress gradient from pith to bark. This itself can result in significant bow and crook deformation when the log is sawed into boards.

The optimal utilisation of a wood raw material is dependent on the wood properties. In this study near infrared (NIR) spectroscopy was used to non-destructively predict density, modulus of elasticity, and modulus of rupture for small clear specimens cut from Picea abies (L.) Karst. (Norway spruce) trees. NIR spectra were recorded directly on the wood surface of each specimen as close as possible to the fracture developed during the bending test. Models were calibrated using partial least squares regression. The validation method was test set validation by data splitting. The correlation between predicted and measured values was highest for modulus of elasticity (0.86), followed by modulus of rupture (0.84) and density (0.79). The validation of the models showed that the average accuracies of predictions were 20.3 kg/m³ for density, 1.1 GPa for modulus of elasticity, and 6.1 MPa for modulus of rupture. Results are comparable to what has been found for other tree species, such as Pinus radiata D. Don (radiata pine), Pinus taeda L. (loblolly pine), and Larix decidua Mill. (European larch).

Models for predicting density, modulus of elasticity, and modulus of rupture based on NIR spectra outperformed simple regression models using the mean annual ring width as the independent variable. NIR spectroscopy is a rapid tool for characterising organic materials. It requires minimal sample preparation and spectra are collected on solid wood, rapidly and nondestructively. For these reasons the method should be tested in production lines for lumber. In future research we aim for reliable predictions of mechanical properties of industrially manufactured lumber using models based on NIR spectroscopy and multivariate statistical methods.

Considerable areas in the northern parts of Norway are afforested with Picea abies (L.) H.Karst. (Norway spruce), Picea x lutzii Little (Lutz spruce), and Picea sitchensis (Bong.) Carrière (Sitka spruce). The species have different machining and wood properties but are similar in visual appearance. We evaluated whether near infrared (NIR) spectroscopy combined with multivariate statistical modelling could be used to identify wood from these three species. In all, 83 wood specimens were available for analyses, 36 of which were used as a test set for model validation. NIR spectra were obtained on the cross-sectional surfaces.

An initial principal component analysis indicated that little information from the first and second components could be used for discrimination, but in score-plots of the third and fourth components the samples from the tree species formed clusters. This showed that the NIR spectra did contain information relevant for tree species identification, and that only a small fraction of the total variance could be used for that purpose.

For classification of the wood specimens, partial least squares discriminant analyses were applied. All 47 specimens in the training set were fitted into the correct group. The test set validated results showed that except for two wood specimens, all specimens were correctly classified. The two misclassified samples were Sitka spruce. This study showed that development of well-performing prediction models for differentiation of wood from Norway spruce, Lutz spruce, and Sitka spruce is possible.

Latent class regression is a statistical method which is not well known in wood science for predicting distortion of sawn timber from structural wood characteristics. The method identifies unknown subgroups in a dataset and allows more accurate regression models to be derived. We identified two separate classes to describe the relationship between bow_dry, spring_dry, or twist_dry and the predictors: initial distortion bow_fresh, spring_fresh, or twist_fresh, wood density, ring width, ring orientation, wood type (juvenile or adult), percentage compression wood measured separately on the four faces, and the contagion index which is a measure for the distribution of compression wood. The latent class regression models developed for the separate classes explained the variation in bow, spring, and twist to a higher degree than a single regression model over the entire dataset. For bow, R² increased from 0.13 to 0.24 and 0.41 for Class 1 and Class 2, for spring from 0.24 to 0.45 and 0.67, and for twist from 0.15 to 0.38 and 0.33. For individual regression models, the predictors showed a varying effect. In classes with significant compression wood on the faces, the effect of wood type seemed weaker, and vice versa. It was concluded that latent class regression analysis allows a more detailed explanation of the effects of wood structure on sawn timber distortion for heterogeneous datasets.

Liquid water sorption in the longitudinal direction in wood samples of Picea abies (L.) Karst. (Norway spruce) was measured with computed tomography (CT) scanning and image processing and then evaluated using multivariate discriminate analysis. The purpose was to determine if there were any differences in liquid water sorption that could be dependent on the vertical position within the tree (0.8, 5.8, and 9.5 in from the butt cut), the growing site (dry or wet), and the type of tree (suppressed or dominant). Test pieces were CT scanned after 1, 3, 7, and 14 days of water sorption in end grain and during desorption at room temperature. The objective was to find wood suited to exterior use that is durable because it takes up water poorly. The conclusion was that heartwood of spruce absorbs less water than sapwood. Heartwood gradients were generally steeper, with a markedly lower moisture content than sapwood. The moisture content gradient profiles differed between the wet and dry sites during sorption and desorption in heartwood and sapwood. Whether or not the trees had been suppressed or dominant had no impact on the moisture content gradients. There was an indication that moisture content gradients in heartwood differed between the first and the second logs, but in sapwood there was no difference.

In recent years there has been growing interest in uneven-aged or mixed forest stands as a result of new demands of society and of changing forestry practices. Unfortunately, due to their complexity, the dynamics of these stands are more difficult to understand than those of pure and even-aged stands. Thus, new research questions have arisen in terms of stand description, stand dynamics, and growth modelling.

The first step toward better management of such mixed or uneven-aged stands is to describe them precisely - this step corresponds to a classical typological approach. As spatial structure plays a key role in the dynamics of such stands, spatial structure analysis can be used to infer certain types of information based on the biological processes involved in the growth and the dynamics of heterogeneous stands, and thus to build a typology.

In order to present a first typology of Quercus petraea (Matt.) Liebl. (Sessile oak) and Pinus sylvestris L. (Scots pine) mixed stands, we analysed the spatial structure in a mixed stand of oak and Scots pine from the French Centre region. We used the classical Ripley function L(r), and intertype function L₁₂(r) to characterise the specific spatial structure of each population, and the structure of the interaction between populations. We then used the results of this analysis to build a typology for these stands, with four main types. These four types may have resulted from ecological processes and historical management, and a typology could be used to simulate realistic virtual stands when real data are unavailable.

A stem and branch growth model, TreeBLOSSIM, has been developed for Pinus radiata D. Don that predicts the location and diameter of branches adjacent to the stem on an annual basis. Research is under way to extend the model to predict wood properties in three dimensions: vertically with increasing tree height, radially with increasing tree age, and circumferentially around a growth ring.

Four studies were carried out to examine the role that branches may have in influencing the 3-dimensional variability of wood properties. These studies illustrated how wood fibres were arranged in the vicinity of abranch, and how the stem cross-sectional shape varied through a cluster of branches; and they indicated that the wood properties in the internode below a branch cluster may be influenced by the diameter of the branches.

Intensive management may adversely affect lumber yield and quality by increasing knot size and creating a more conical stem form with a greater average rate of taper. This study was initiated to examine the impact of management on simulated lumber yield and quality. Stem diameter and branch size and location of 223 Pseudotsuga menziesii (Mirb.) Franco (Douglas fir) stems ranging in age from 5 to 65 years and from a wide variety of stand conditions were intensively measured. Stand conditions included varying levels of vegetation management, precommercial thinning, commercial thinning, fertiliser application, and severity of infection by Phaeoctyptopus gaeumannii (Rohde) Petrak (Swiss needle cast). In addition, 86 virtual logs were created and processed by AUTOSAW. Significant changes in both stem form and branch characteristics were observed among the stand conditions examined, with maximum branch size being the most responsive to silvicultural regime and disease severity. Changes related to fertiliser and thinning were not significant enough to adversely affect simulated lumber quality and yield. Indices of branch size were poor predictors of simulated log grade yield. Although quantification of branch size and location is important for understanding crown structure, growth potential, and the vertical distribution of biomass, factors such as juvenile wood percentage and wood density may exert more control over simulated product quality in the young Douglas fir analysed in this study.

Process-based models have advanced to a level which enables their utilisation for evaluating forest management options, as well as their use in environmental education. To date, such applications are relatively rare, but they could be promoted by means of appropriate interactive and user-oriented interfaces for model simulation.

The PuMe-software has been constructed as an interactive tool for forestry studies at secondary and university levels. The objective was to build an interactive and user-oriented interface for running a forest growth model with either user-designed or pre-determined inputs of forest management options such as stocking densities, thinning practices, site fertilities, and fertiliser application. The PipeQual model was chosen as the growth simulator because of its versatility in predicting traditional forestry characteristics, stem structure (stem shape, knot zones, heartwood, and sapwood), biomass, and carbon balance.

The first version of PuMe is now in test-use in forestry education by universities and secondary schools, and by individual forest owners. However, it is applicable only to Pinus sylvestris L. (Scots pine) under a limited set of conditions. In an ongoing project, PuMe II, the software will be developed further by adding various new features: simulation of spruce growth, forest damage, fertiliser application, pruning, and more flexible thinning methods. Especially, a new module will be incorporated to allow for a more detailed visualisation of stem properties (growth rings, wood density, fibre properties).

Oak is the most important species in French forestry. This wood is purchased by a great variety of firms, with prices that can range on a scale from 1 to 100. The main determinant of wood use is its quality and this depends on the presence or absence of defects in the wood. There is a great diversity of defects, but their importance for estimation of wood quality is not well-known. We have formulated a methodology for measuring the severity of defects in the final use of the wood. The measurement is based on two complementary indicators. The first one evaluated the drop in price caused by the presence of defects on individual logs sold during German sales of felled logs. As prices are linked not only to the severity of defects and as we did not come across a complete range of users or defects during this sale, a second indicator was constructed to complement the first one. This second indicator was based on a survey in which people were asked to sort virtual logs in order of preference. This methodology is useful for an initial approach to the problem but some improvements are needed to provide complete answers.

Predictions of the branching characteristics of Sitka spruce (Picea sitchensis (Bong.) Carr.) were made as part of the development of a timber properties simulation tool. For each annual growth unit, non-linear modelling was used to describe the average number of branches and their associated diameter, insertion angle, and probability of being alive. The parameters for the model were obtained using the branching characteristics of 60 trees collected from a range of locations across Scotland and northern England. The non-linear equations describing the branching properties gave an adequate representation of the mean of each branch property from the top to the base of the tree. Analysis of the residuals around the stem revealed that branches grew bigger on the south side but were more numerous on the north side. The predictions used a total of only six predictor variables that can be obtained from normal mensurational data and from the height growth history of the trees. Yield tables could hence be used to run two simulations of typical Sitka spruce stands managed under a thinning or no-thinning regime. As expected, the effect of thinning was to increase branch diameter, to lower the height at which branches die, and to increase the number of branches without having any effect on their insertion angle. After incorporation of clear-wood properties, the model will be used to predict the properties of sawn battens.

Clearwood conversions were linked to stand growth using a multiple regression approach. Stand growth of Pinus radiata D. Don was modelled for a range of treatments that spanned extremes in site productivity, stocking levels, pruning practices, resin pocket incidence, and genetic seed sources. At 20 years and thereafter annually until 30 years of age, a pruned butt log was theoretically cut from the mean stem and sawn, and clearwood timber grades were calculated and expressed as a percentage of sawn timber volume. A multiple regression model that linked rotation age to clearwood conversions, site productivity, and treatments was developed for GF22 seedlot.

Application of the model to a site of moderate productivity, planted at 800 stems/ha, thinned to 400 stems/ha, and pruned when diameter over stubs measured 180 mm, indicated that a rotation of at least 33 years would be needed to achieve a clearwood production level of 40%. On extremely low or high productivity sites, rotations of 40 and 25 years respectively were predicted. At a final stocking of 200 stems/ha and under the same practices, rotations of 32, 25, and 17 years were predicted for the low, moderate, and high productivity sites respectively.

Preliminary validation of the model, through comparison of predicted age with that of real logs ranked as "stars" and "super-stars", according to processing efficiency and profitability, produced close correspondence in rotation age. The regression model predicted rotation ages for stars and super-stars to be 27.5 and 33.0 years respectively while actual ages were 28.1 and 32.0 years respectively.

A model for automatic bucking of sound-knot sawlogs has been implemented in Timberjack and Ponsse cut-to-length harvesters. It is based on a relationship between diameter at breast height (dbh) and the largest small-end diameter (s.e.d.) of logs to produce sound-knot sawnwood in the centre boards. In order to evaluate the possibilities for using this new functionality of harvesters in practical applications, a series of studies were carried out to identify, buck, and sort logs for a commercial appearance-grade of Pinus sylvestris L. (Scots pine) marketed for furniture industries. In an initial calibration study, the lowest in-grade height-level was determined for taper-sawn logs from seven stands of different ages and growth rates in central Sweden. The mean soundknot quotient (SKQ, calculated as small-end diameter at this threshold limit divided by diameter at breast height) was 0.73 (0.69 for final cutting and 0.78 for thinnings). Using these data, the effects of different parameter settings were simulated to evaluate automatic classification of logs. It was apparent that a restrictive setting (a low sound-knot quotient-level forcing the bucking limit higher up the stem) could lead to a high proportion of correctly classified in-grade logs, but at the cost of missing furniture raw material through incorrect classification of logs as out-of-grade. Three sound-knot quotient-levels (0.64, 0.68, and 0.72) were tested in a practical study to evaluate log sorting in two final cutting stands and for two lumber dimensions (50 × 125 mm and 50 × 150 mm). For each setting and stand, logs with s.e.d. 192-239 mm were automatically selected by the on-board computer, and identified (in-grade, out-of-grade, or butt-log) with the paint-spraying function of the harvester head. The results verified the findings of the simulation study: the proportion of correctly classified logs (success rate) changed considerably with the different settings, and a high success rate also resulted in a substantial number of out-of-grade logs (as classified by the harvester) containing furniture-grade sawnwood. Finally, a large-scale, industrial trial was performed using SKQ 0.70 and a normal production environment involving several harvesters and a large sawmill. In total, of the 1087 logs classified as in-grade by the harvester computer, 86% fulfilled the furniture grade requirements for sawnwood when converted in the sawmill. Correspondingly, of the 622 non-butt logs classified as out-of grade, 63% did not meet the furniture grade requirements.

During pre-harvest assessment, wood quality information is measured with the software package ATLAS Cruiser. Various modelling methods are applied to this information to predict log yields for different stands, through time and under different cutting strategies. This yield information allows the scheduling component (ATLAS Market Supply) to find an optimal way of meeting the demand for logs of varying qualities from the available stands, given harvesting capabilities and transport distances.

Assuming the issues arising from initial use of the system can be addressed, the indications are that the combined system of pre-harvest forest assessment and optimisation of harvesting and log allocation is able to characterise the resource in terms of wood qualities and provide the best match of logs to customer demands, thereby maximising the returns to both the grower and the processor.

Since the 4th IUFRO WP 5.01.04 Workshop held in Canada in 2002, our research team, the LERFoB, in co-operation with two other French organisations, CTBA and ONF, has made a new advance on a long-term project aiming to optimise the chain from the plant to the plank in sessile oak (Quercus petraea Liebl.) taking into account considerations related to sustainable management. The method used in this project is typical of the method in use at the LERFoB - that is, a method based on joint modelling of growth and wood quality as well as simulation software. There are nine elements in our chain of models and software. These are now available to perform detailed simulations of the effect of contrasted silvicultural schedules - in a context of pure and even-aged stands naturally or artificially regenerated - on products delivered, environmental considerations, economical considerations, and employment.

These proceedings contain 17 papers on the forestry conversion chain and modifications of silvicultural practices and schedules. Different techniques and computer models used in differentiation of forest products as well as assessment of wood properties, quality and value are also discussed.

Tracheary element differentiation and secondary cell-wall formation have been studied in cell cultures of coniferous gymnosperms. Factors that influence tracheary element differentiation are (i) the sucrose concentration in the media, (ii) the concentration of nutrients in the media, (iii) temperature and light, and (iv) the types and concentrations of phytohormones in the media. There are advantages and disadvantages in using cell cultures for studying tracheary element differentiation and secondary cell-wall formation, but in combination with in planta studies the cell culture approach is very useful for advancing our understanding of these processes.

There are techniques are available for the analysis of plant cell wall structures that are fast and require small amounts of material. Some of these approaches include oligosaccharide mass profiling, carbohydrate gel electrophoresis, and capillary electrophoresis.

Flowers are prized as objects of great beauty and diversity, and are commercially valuable (~US$4.5 billion in international trade yearly) and highly perishable. Biologically, flower petals have an important role in the lifecycle of plants, as they protect immature reproductive structures, then provide the attraction and accessibility needed for pollination to occur. Flower petal cell wall construction, maintenance, andbreakdown are important factors in the life of flowering plants as well as being related to the visual quality of commercial cutflowers. The petal is generally a thin structure, with a mesophyll-type cell layer between an ordered upper and lower epidermis. Petal cell walls are constructed in such a way as to be able to accommodate large and quite fast increases in cellular turgor during petal expansion and flower opening. There is some variety in the way the flower lifecycle may end once pollination has occurred, with petal wilting, shattering, abscission, almost complete autolysis of petal tissue, and the development of a papery shell all occurring in different species. The senescence of flower petals can be accompanied by increased activity of cell wall hydrolases, similar to that found during fruit ripening, and degradation of pectic and hemicellulosic polysaccharides. For some flowers, however, wall dissolution is restricted to depolymerisation of hemicelluloses and loss of neutral sugars, particularly galactose and arabinose. The few species in which the metabolism of cell wall polysaccharides in flower petals has been studied in depth include carnation, sandersonia, and daylily.

Kiwifruit is an excellent model in which to study events in the cell wall that lead to fruit becoming soft. In kiwifruit, softening can be separated into distinct phases, with the phase associated with the most extensive and rapid loss of firmness being well-separated temporally from that associated with the respiratory climacteric and ethylene production. This contrasts with tomato, avocado, and many other fruit where softening changes, ethylene production, and the climacteric occur concurrently. The changes that occur in the kiwifruit cell wall during the softening process have been extensively characterised by chemical analyses, by histochemical and immunolocalisation techniques, and by monitoring the activity and expression of wall-associated enzymes.

Fruit softening is an important part of the ripening process, and involves changes to cell turgor and primary cell wall structure. Most of the polysaccharide components of the cell wall are subjected to some degree of controlled degradation, resulting in a loosening and swelling of the wall structure, a weakening of cell wall strength, and reduced intercellular adhesion. Early ripening changes involve the degradation of the galactan/arabinan side chains of rhamnogalacturonan-I, demethylesterification of homogalacturonan, and depolymerisation of matrix glycans (hemicelluloses). Solubilisation of pectins increases during ripening, but depolymerisation of pectins is usually most pronounced late in ripening. Considerable variation in the extent of pectin depolymerisation and galactan/arabinan loss exists between species. Transgenic studies have shown that expansin may control cell wall loosening, and b-galactosidase may be important in increasing cell wall porosity. Suppression of either of these enzymes resulted in a retention of fruit firmness. Suppression of endo-polygalacturonase and pectin methylesterase had little effect on fruit firmness during ripening, but influenced fruit shelf life due to alterations in the integrity of the middle lamella, which affected intercellular adhesion. The enzyme(s) responsible for depolymerisation of matrix glycans have not been defined, and the identity of the ripening-related xyloglucanase remains obscure.

Chemical modification with chitin- and chitosan-hexamethyl methylol melamine (HMMM) co-polymers was investigated for improving the stiffness of lignocellulosic materials. Chitin and chitosan were converted by chemical means to low molecular weight oligosaccharides with molecular weight profiles suitable for penetration of lignocellulosic cell walls. The oligomers were reacted under controlled conditions with hexamethyl methylol melamine (HMMM) to produce aqueous formulations of oligosaccharide bonded to HMMM, the "pre-polymers". The chitosan oligomers reacted with HMMM to produce, on condensation polymerisation, a water-insoluble polymer in high yield (69%), whereas the chitin oligomer HMMM condensation reaction gave poor co-polymer yields (28-34%). The yield of co-polymer from the condensation polymerisation reaction was critical to the success of the cell wall modification in improving stiffness.

Pinus radiata D. Don veneers were treated with chitin and chitosan oligomer HMMM formulations to average dry weight percentage gains of 69% and 57% respectively. No improvement in veneer stiffness was obtained with the chitin oligomer HMMM treatment, whereas the chitosan oligomer HMMM treatment resulted in an average veneer stiffness enhancement of 20%. There was a linear relationship between the level of stiffness improvement and the degree of co-polymerisation of the oligomers with HMMM. A threshold of greater than approx. 30% co-polymer yield was necessary before any improvement in veneer stiffness was observed. Polysaccharides with a b-(1->4) configuration, such as chitosan, therefore offer potential for lignocellulosic stiffness property modification.

The first example of modified wall architecture in woody cells examined was in poplar trees where wounding caused wall thickenings of xylem fibres differentiating at the time of wounding. These fibres close to a wound also displayed slightly higher lignin content and an inhomogeneous lignin distribution as revealed by UV-microspectrophotometry. Additionally, the lignin in the middle lamella and the S₂ layer of modified fibres contained more guaiacyl units than fibres of normal wood. These wound response mechanisms are assumed to contribute to an increased resistance. Secondly, in a pine seedling displaying stem bending, extensive compression wood formation was observed. Autofluorescence confirmed that the lignin content in the compression wood tracheids was distinctly higher than in normal tracheids. Immunolabelling of galactan for the first time clearly demonstrated that the bulk of galactan is localised in the outer S2 wall regions.

Confocal Raman microscopy was used for chemical imaging of wood cell wall polymers and to follow molecular changes during tensile deformation. Spectral maps were acquired from cross-sections of poplar wood and images calculated by integrating the intensity of characteristic spectral bands. This enabled direct visualisation of the spatial variation of the lignin content without any chemical treatment or staining of the cell wall. A higher lignin content was visualised in the cell corners, the compound middle lamella, and the secondary cell wall of vessels than in the fibres. The S1 was distinguished from the S2 by integrating over the band at 1097 cm^-1, because the intensity of this vibration is sensitive to the orientation of the cellulose molecule. The position of this band was shifted towards shorter wavenumbers during straining of wettangential sections, demonstrating that the cellulose molecule was subjected to a deformation. The band shift was followed during the tensile test and a good trend and correlation with strain and stress were observed. Investigating tissue types with different properties and cell wall assemblies will help to reveal the polymer composition and orientation non-destructively with a high spatial resolution. By investigating structural changes during tensile straining, we aim at understanding the different stress-strain behaviour and the molecular mechanistic phenomena involved.

Fourier transform infrared (FTIR) imaging offers the opportunity to analyse the chemical composition of wood spatially resolved. To illustrate the potential of FTIR imaging for wood analysis, the lignin distribution in cross sections of beech and poplar wood was analysed and the spatial resolution compared. Additionally, the ratio of guaiacyl/ syringyl lignin in a poplar wood section was computed. The resolution of the FTIR microscope was sufficient to resolve individual cell walls of poplar wood fibres.

Light and transmission electron microscopy are used in studying wall morphology and histochemical methods, including immunocytochemistry, can be used to locate specific compounds in walls. All plant cell walls contain a fibrillar phase of cellulose microfibrils and a matrix phase which contains a high proportion of non-cellulosic polysaccharides that vary in their chemical structures, depending on wall type and plant taxon. The non-cellulosic polysaccharide compositions of three common wall types - lignified secondary walls, non-lignified secondary walls, and non-lignified primary walls - exemplify this. The principles used in constructing the most recent models of non-lignified primary walls can be used in modelling lignified secondary walls.

This special issue contains 10 papers providing information on various aspects of primary and secondary cell walls of vascular plants particularly on cell wall modification, biological development, and the different methods of cell wall analysis of wood.

Measuring the height of a tree takes longer than measuring its diameter at breast height and often the heights of only a subset of trees of known diameter are measured in forest inventories. Accurate height-diameter equations must therefore be used to predict the heights of the remaining trees. Two trees within the same stand and that have the same diameter are not necessarily of the same height; therefore we developed a deterministic equation, using the Schnute function, and then added a stochastic component to it, to mimic the real natural variability in height. The stochastic approach uses the standard error of a new observation in a similar way to the method of obtaining the prediction interval for an individual (new) in a regression model, but rather than using the t value corresponding to a fixed limit for all the trees, it uses a pseudo-random number having a normal distribution N(0,1) for each observation. The stochastic approach was evaluated with data from four thinning trials located in single-species, even-aged stands of the most commercially important pines in Galicia (north-western Spain). More realistic height predictions were obtained than with the deterministic model for individual diameter classes, as demonstrated by the results of the Kolmogorov Smirnov test and by visual analysis of box plot graphs.

The host range and geographic distribution of the basidiomycete fungus Chondrostereum purpureum (Pers.) Pouzar in New Zealand were determined through analysis of herbarium records from Landcare Research and the New Zealand Forest Research Institute Limited, as well as published reports. The fungus has been recorded in every geographic region of the North Island, with the exception of Northland and Rangitikei, and from the northern portion of the South Island, as well as Southland, Otago Lakes, south Canterbury, and mid Canterbury, but it is known to be present throughout New Zealand. It has been recorded on 23 angiosperm families and 1 gymnosperm family in New Zealand. Based on the geographic distribution and epidemiological studies of the pathogen that have been conducted elsewhere, it is concluded that the utilisation of C. purpureum as an inundative biological control agent would not significantly alter the risk of infection by C. purpureum within New Zealand.

The five fungi described in this paper have been recorded from New Zealand but not fully described. The fungi are:

Corticolous Ascomycota: Cosmospora sp. (aff. Cosmospora purtonii (Greville) Rossman & Samuels) on Myrsine australis (A.Richard) Allan.

Caulicolous Ascomycota: Colpoma quercinum (Persoon) Wallroth on Quercus sp.; Hysterographium fraxini (Persoon) De Notaris on Fraxinus spp.

Foliicolous Coelomycetes: Pestalotiopsis karstenii (Saccardo & P. Sydow) Steyaert on Banksia sp.

Staninwardia breviuscula B. Sutton on Eucalyptus spp. and Metrosideros excelsa Solander ex Gaertner.